A System and Method for Detection of Deviations in Packaging Containers

1. A method for detection of deviations in packaging containers for liquid food produced in a machine, the method comprising: creating a virtual model of a packaging container in a virtual coordinate system (x, y, z), defining a deformation zone on a surface of the virtual model, creating a defined deviation in the deformation zone having a defined geometry and coordinates in the virtual coordinate system to create a controlled deformation of the virtual model, wherein the defined deviation is defined as a concave and/or convex shape in the surface of the virtual model, producing an image rendering of the virtual model with said controlled deformation to generate image features representing a deviation in the packaging container, associating the image features with different categories of deviations, and inputting the image features to a machine learning-based model for subsequent detection of categories of deviations in packaging containers in the machine based on the image features, the method further comprising: defining coordinates of a deformation line extending along the surface, and folding the surface along the deformation line with a defined angle to form the concave and/or convex shape in the surface.

2. The method according to claim 1 , further comprising mapping a décor of defined color and/or pattern on the surface of the virtual model.

3. The method according to claim 1 , wherein producing the image rendering comprises processing the virtual model according to defined lighting conditions.

4. The method according to claim 1 , further comprising defining a surface roughness on the surface of the virtual model.

5. The method according to claim 1 , comprising defining a pattern on the surface by: mapping a two-dimensional image of the pattern on the surface, wherein the two-dimensional image comprises image data that defines a gradient, and changing, at a position of the pattern, the orientation of the surface so that a surface normal of the surface is changed according to the gradient.

6. The method according to claim 5 , wherein the pattern comprises least one crease line.

7. The method according to claim 1 , wherein creating the virtual model comprises: defining a geometry of a blank for the packaging container in the virtual coordinate system, defining crease lines of the blank, and folding the blank along the crease lines in the virtual coordinate system to create the virtual model.

8. The method according to claim 1 , further comprising defining a virtual camera position in the virtual coordinate system in relation to the virtual model so that a viewpoint from which the image features are obtained in the image rendering corresponds to a viewpoint from a camera position for said subsequent detection in a coordinate system of the machine.

9. The method according to claim 1 , further comprising detecting with the machine learning-based model one or more deviations in another packaging container being processed by the machine based on image features of the another packaging container.

10. A non-transitory storage medium storing a computer program which, when executed by a computer, causes the computer to carry out the steps of the method according to claim 1 .

11. A system for deviation detection in packaging containers for liquid food produced in a machine, the system comprising: a processing unit configured to: create a virtual model of a packaging container in a virtual coordinate system (x, y, z), define a deformation zone on a surface of the virtual model, create a defined deviation in the deformation zone having a defined geometry and coordinates in the virtual coordinate system to create a controlled deformation of the virtual model, produce an image rendering of the virtual model with the controlled deformation to generate image features representing a deviation in the packaging container, associate the image features with different categories of deviations, and input the image features to a machine learning-based model for subsequent detection of categories of deviations in packaging containers in the machine based on the image features, wherein the processing unit is further configured to define a pattern on the surface by mapping a two-dimensional image of the pattern on the surface, and wherein the image comprises image data that defines a gradient, and wherein, at a position of the pattern, the processing unit is configured to change the orientation of the surface so that a surface normal of the surface is changed according to the gradient.

12. The system according to claim 11 , wherein the processing unit is further configured to map a décor of defined color and/or pattern on the surface of the virtual model.

13. The system according to claim 11 , wherein the processing unit is further configured to detect with the machine learning-based model one or more deviations in another packaging container being processed by the machine based on image features of the another packaging container.

14. A method for detection of deviations in packaging containers for liquid food produced in a machine, the method comprising: creating a virtual model of a packaging container in a virtual coordinate system (x, y, z), defining a deformation zone on a surface of the virtual model, creating a defined deviation in the deformation zone having a defined geometry and coordinates in the virtual coordinate system to create a controlled deformation of the virtual model, producing an image rendering of the virtual model with said controlled deformation to generate image features representing a deviation in the packaging container, associating the image features with different categories of deviations, and inputting the image features to a machine learning-based model for subsequent detection of categories of deviations in packaging containers in the machine based on the image features, the method further comprising defining a pattern on the surface by: mapping a two-dimensional image of the pattern on the surface, wherein the two-dimensional image comprises image data that defines a gradient, and changing, at a position of the pattern, the orientation of the surface so that a surface normal of the surface is changed according to the gradient.

15. The method according to claim 14 , wherein the pattern comprises least one crease line.

16. The method according to claim 14 , further comprising mapping a décor of defined color and/or pattern on the surface of the virtual model.

17. The method according to claim 14 , wherein producing the image rendering comprises processing the virtual model according to defined lighting conditions.

18. The method according to claim 14 , wherein creating the virtual model comprises: defining a geometry of a blank for the packaging container in the virtual coordinate system, defining crease lines of the blank, and folding the blank along the crease lines in the virtual coordinate system to create the virtual model.

19. The method according to claim 14 , further comprising defining a virtual camera position in the virtual coordinate system in relation to the virtual model so that a viewpoint from which the image features are obtained in the image rendering corresponds to a viewpoint from a camera position for said subsequent detection in a coordinate system of the machine.

20. The method according to claim 14 , further comprising detecting with the machine learning-based model one or more deviations in another packaging container being processed by the machine based on image features of the another packaging container.